Eighteen year (1985-2002) analysis of incidence, mortality, and cardiac procedure outcomes of acute myocardial infarction in patients > or = 65 years of age.

The temporal patterns of outcomes and therapy in patients aged > or =65 years with acute myocardial infarctions (AMIs) from a national database were examined to better understand this increasingly important demographic group. The National Hospital Discharge Survey (NHDS), a nationally representative sample of acute care hospitals in the United States, was used for analysis. Using International Classification of Diseases, Ninth Revision, Clinical Modification codes, hospitalizations with first-listed diagnoses of AMI from 1985 to 2002 were identified. A multivariate logistic regression model was developed to identify predictors of mortality in these patients. The logit of propensity score was used as an adjuster for reducing the bias of nonrandom assignment of cardiovascular procedures. Although the number of patients aged > or =65 years admitted with AMIs increased over the study period, the incidence in patients aged > or =65 years decreased by 1.5%. In patients aged <65 years, the incidence was significantly lower for any year and decreased by 17.5%. The in-hospital mortality rate in patients aged > or =65 years decreased (from 22.0% in 1985 to 11.5% in 2002) but remained significantly higher compared with younger patients (from 7.0% in 1985 to 3.2% in 2002). Overall cardiac procedure use increased from 18.0% in 1985 to 50.7% in 2002, but patients aged > or =65 years consistently underwent fewer procedures than younger patients. After taking 12 covariates into consideration, not undergoing a cardiac procedure remained a significant risk factor for mortality (odds ratio 3.13, p <0.0001). Acute renal failure (odds ratio 4.64, p <0.0001) and age > or =65 years (odds ratio 2.14, p <0.0001) were the other 2 strongest independent predictors of mortality. In conclusion, the incidence of AMI is decreasing. Over the 18-year period from 1985 to 2002, there was a significant reduction in mortality, but patients aged > or =65 years remained at particular risk. Multivariate analysis suggests that a lack of the use of procedures in these patients may at least partially explain their higher mortality.